Evaluating Time Synchronization Performance on an Edge Device Using Geographically Distributed NTP Servers

Time synchronization is essential for Internet of Things (IoT) and edge computing applications that require accurate time stamping and coordinated operation. Low-cost edge devices employ quartz-based oscillators, which exhibit clock drift over time and Network Time Protocol (NTP) is widely used to mitigate this kind of drift and synchronized to the reference clock. However, achievable NTP synchronization performance is strongly influenced by network conditions and the geographic location of the reference time server. This article presents a study over long period of NTP-based time synchronization on a low-cost edge device, i.e., Raspberry Pi using chrony and geographically distributed NTP servers. Clock offsets and round-trip network delays are recorded at one-minute intervals using chrony tracking statistics, resulting in independent datasets for each server. The collected data are analyzed for evaluating synchronization accuracy, stability, and client-side synchronization uncertainty under realistic network conditions. Client-side uncertainty evaluation has been done for the practical measure of synchronization reliability as experienced by an edge device. The uncertainty analysis also offers a realistic basis for relative performance assessment and informed NTP server selection in IoT and edge computing deployments.